Predicting the impact of diet on the human intestinal microbiota

Transcription

1 Predicting the impact of diet on the human intestinal microbiota Sylvia H. Duncan and Harry J. Flint Microbiology Group, Rowett Institute of Nutrition and Health, University of Aberdeen, UK ROWETT INSTITUTE OF NUTRITION AND HEALTH

2 Impact of the gut microbiota on health metabolism of dietary components immune function, inflammation modification of host secretions (mucin, bile, gut receptors..) energy, nutrient supply Gut function, gut disorder Infection, irritable bowel syndrome, colitis, colorectal cancer pathogenesis barrier function Systemic health Energy supply, satiety Diabetes Heart disease Autoimmune disorders

3 Microbially derived metabolites Damaging Protective Indoles Anti-Inflammatory Molecules Heterocyclic Amines N N N H 2 N Bile Acids O H N O O H H O Phytoestrogens O H O O H O O H O O H N-Nitrosamines O N N H O N H 2 H H H H Polyamines N Short Chain Fatty Acids N H 2 O O H Anti-oxidant Molecules H O O O H O H O H O H

4 Human gut microbiota Factors that impact on bacterial persistence in the colon Host factors -Microbiota acquired at birth; aging -Host immune response

5 Changes in the composition of the human colonic microbiota with aging Proteobacteria E. coli F. prausnitzii [Duncan and Flint, Maturitas 2013]

6 Human gut microbiota Factors that impact on bacterial persistence in the colon Host factors -Microbiota acquired at birth; aging -Host immune response Gut environmental factors nutrient availability/diet macronutrient and micronutrient availability ph bile gut transit (wash out) anaerobiosis

7 Interplay between diet and microbiota on gut health Protection against colorectal cancer and colitis Nutrient supply to mucosa Barrier against infection Release of phytochemicals Interplay between diet and microbiota Exposure to metabolites and bacteria that promote disease phenols, amines, indoles, N-nitroso compounds, H 2 S, amines, bile acids, faecapentaenes, heme

8 Principal substrates available for utilization by intestinal microbes Of dietary & intestinal origin: range Resistant starch Non-starch polysaccharides Unabsorbed sugars Oligosaccharides Dietary protein Enzymes / secretions / mucus Amount [gram per day] 50 [from Cummings & Macfarlane (1991)]

9 Fermentation of dietary macronutrients in the large intestine Acetate Propionate Butyrate Absorption / g gut contents (large intestine) Short Chain Fatty Acids + H 2 + CO 2 + CH 4 Outnumber human cells in the body by 10:1 fermentation Dietary Polysaccharides Several hundred bacterial species colonise each individual Most are oxygen sensitive, but can be cultured

10 Overview of the microbiology of the human large intestine Several hundred bacterial species inhabit the large intestine Range Bacterial Groups Bacteroidetes Clostridial cluster IV Clostridial cluster XIVa Other clostridial clusters Actinobacteria Proteobacteria Verrucomicrobia Approx. 99 % of colonic bacteria belong to 4 phyla Bacteroidetes Firmicutes Actinobacteria Proteobacteria Proportion of total bacteria (%) Red bars indicate Gram ve bacterial groups and blue bars for Gram +ve groups

11 Dominant species of human colonic bacteria (26 faecal samples - 6 overweight male volunteers) Other 50% phylotypes (72% uncultured) 25 cultured species accounted for approximately 50% of 16S rrna sequences Faecalibacterium prausnitzii Eubacterium rectale Colinsella aerofaciens Clostridium clostridioforme Bacteroides vulgatus Anaerostipes hadrus Ruminococcus bromii Eubacterium hallii Blautia wexleri Bacteroides dorei Roseburia faecis Dorea longicatena Subdoligranulum variabile Bacteroides uniformis Ruminococcus obeum Bacteroides ovatus Blautia luti Parabacteroides distasonis sp nov A2-166 sp nov SR1/5 Lachnospira pectinoschiza sp nov 80/3 Dialister invisus Roseburia inulinivorans Ruminococcus callidus others Walker AW et al ISME J (2011) Flint HJ et al Gut Microbes (2012)

12 16S rrna sequence analysis of faecal samples 6 volunteers x 4 diets x approx 250 clones per library = 5,920 clones 10 dominant phylotypes phylum % clones Faecalibacterium prausnitzii * Firmicutes (IV) 7.98% Eubacterium rectale* Firmicutes (XIVa) 4.43% Clostridium clostridioforme Firmicutes (XIVa) 3.83% Collinsella aerofaciens Actinobacteria 3.67% Bacteroides vulgatus Bacteroidetes 3.21% Anaerostipes hadrus* # Firmicutes (XIVa) 2.25% Ruminococcus bromii Firmicutes (IV) 2.11% Eubacterium hallii* Firmicutes (XIVa) 2.00% Blautia wexleri Firmicutes (XIVa) 1.89% Bacteroides dorei Bacteroidetes 1.67% * = butyrate producers Sum 33.0% # =new species [Allen-Vercoe, et al. 2012]

13 Butyrate formation from lactate and acetate by Eubacterium hallii and Anaerostipes spp. 4 lactate ADP ATP acetyl- P 2 H 2 2 Fd 2 NADH 2 2 NAD 4 Fd acetate 8 [H] 2 FdH 2 2 FdH 2 4 pyruvate 4 acetyl-coa + 2 acetyl-coa 3 CoA 6 [H] 3 H 2 O 6 [H] 3 butyryl-coa 4 CoA 4 CO 2 2 acetate Concentration (mm) Lactate accumulation in stool samples of colitis patients (Vernia et al., 1988) Controls Quiescent colitis Mild colitis Moderate colitis Severe colitis CoA P i acetyl-coa 3 butyrate 4mols lactate+2 mols acetate 3 mols butyrate+4 mols CO 2 [Duncan et al, 2004]

14 Selectivity of prebiotics for promoting growth of human colonic anaerobes in pure culture (YCFA medium) phylum species & strain dahlia potato no CHO scfos HP inulin inulin starch XOS Firmicutes F. prausnitzii A2-165 Roseburia faecis M72/1 R. intestinalis L1-82 R. inulinivorans A2-194 R. hominis A2-183 Eubacterium rectale A1-86 Eubacterium hallii L2-7 Anaerostipes caccae L1-92 Coprococcus comes A2-232 C. eutactus L2-50 Actinobacteria Bifidobacterium longum 8809 B. longum B. infantis Bacteroidetes Bacteroides vulgatus B1447 B. thetaiotaomicron 5482 Butyrateproducers OD 650 < 1.0 > 0.4 OD 650 > 1.0 FOS = fructo-oligosaccharides XOS = xylo-oligosaccharides [Scott et al FEMS Microbiol Ecol (2014) - maximal ODs attained in 24h; 0.5% carbohydrate]

15 Impact of low carbohydrate weight loss diets-human studies Amount of dietary carbohydrate in the diet 18 human male volunteers; (av. Age 37 y); (av. BMI 35 kg/m2) M moderate CHO low CHO M low CHO moderate CHO 1 week 4 weeks 4 weeks

16 Impact of diet on microbial metabolites in obese human subjects 18 male volunteers (av. Age 37 y); obese (av. BMI 35 kg/m2) Carbohydrate g NSP g Starch g Protein g Fat g High carbohydrate Moderate carbohydrate Low carbohydrate

17 Response of the major phylogenetic groups of human gut bacteria to dietary change Diet Carbohydrate Total SCFA Butyrate Butyrate (g) (mm) (mm) (%) High CHO Moderate CHO Low CHO % Eub338 or Eub338 /g * % Eubacterial (Eub338) count in faeces * *** * *** High Moderate Low *** 5 0 Bac303 Fprau645 Rfla729 + Rbro730 Bif164 Erec482 Bacterial group (probe) Prop853 Rrec584 Erec - Rrec log 10 total *** P < * P< 0.05 [Duncan, S.H, et al., AEM]

18 Relationship between faecal butyrate concentration and bacteria detectable in faeces with the Rrec584 probe in obese human subjects Diet M normal (high carb);; NK high protein, medium carb; K high protein, low carb Correlation 0.68 (P<0.001) Butyr BuK BuM BuNK regression log10 Rrec

19 Impact of diets upon bacterial fermentation products Comparison of maintenance diet with week 4 mean values for :- medium carbohydrate high protein, low carbohydrate high protein diets Significant increases in branch chain fatty acids on high protein diets from amino acid breakdown (leucine, valine, isoleucine) % total SCFA * * propionate butyrate * valerate * * isovalerate isobutyrate * M MC LC * p < 0.005

20 Impact of high protein low carbohydrate diet 3000 Increase of N-nitrosocompounds on highprotein diet: ATNC (ug/kg) [P<0.001] maintenance medium carbohydrate low carbohydrate Low carbohydrate high-protein diets may increase the risk of disease increase in genotoxic metabolites

21 Major fibre derived phenolics in faecal samples Major fibre-associated phenolics 8 volunteers Concentration (ug cm-3) (ferulic acid derivatives) P < 0.05 P < P < 0.05 Maintenance Diet Low Carb. Diet 20+/-1 days Low Carb. Diet 27+/-1 days 2 Ferulic acid 3OMe4OHPPA 3,4OHPPA 3OHPPA hydrogenation demethylation dehydroxylation HO O H O O H O O H O O O O H O O H HO O H H O [Russell et al, 2011]

22 Effect of inulin on the faecal microbiota of human volunteers Type of dietary carbohydrate Human volunteer cross-over study, 6 volunteers per group: INULIN INULIN Faeces: microbiota changes (real-time PCR) 50 CONTROL CONTROL 21 DAYS 21 DAYS % of universal 16S rrna gene copies Cluster XIVa Cluster IV background control inulin P = P = Bact Erec Rrec2 Rrec1 Ehal Clep Fprau Rum CoAT Bifids [Ramirez-Farias et al. 2009]

23 Resistant starch compared to non-starch polysaccharide diet 14 male volunteers with metabolic syndrome (mean age 54 years, mean BMI 39.4 kg/m 2 ) M: maintenance diet NSP: high non-starch polysaccharide, low RS RS: high resistant starch, low NSP HPMC: high protein, moderate carbohydrate Mean dietary intake [g/d]: Collection of faeces M NSP RS HPMC 1 wk 3 wks 3 wks 3 wks M RS NSP HPMC Diet CHO starch RS NSP protein fat M NSP RS HPMC Weight maintenance Weight loss CHO: carbohydrate added wheat bran added type III resistant starch Walker et al. 2011, ISME J

24 Impact of dietary non-digestible carbohydrate Human volunteer trial 14 obese males Impact on the gut microbiota :- [+ P <0.001] [Walker A.W. et al (2011) ISME J 5, ] M Wheat bran Res Starch Weight loss (n =7) M Res Starch Wheat bran Weight loss (n =7) 1 week 3 weeks 3 weeks 3 weeks

25 Cluster IV Ruminococcus spp. % of universal 16S rrna gene copies M NSP N RRS WHPMC S S L P % starch digestibility Time (d) 0 Volunteer volunteer M RS NSP HPMC Time (d) RS diet NSP diet [Walker et al. ISME J 2011] Volunteer

26 Ability of Ruminococcus bromii to restore resistant starch degradation 100% Volunteer 25 % of residual starch 80% 60% 40% no addition + B. adolescentis + E. rectale + B. thetaiotaomicron 20% mean level of controls + R. bromii 0% Time (h) R. bromii is a keystone species essential for extensive RS fermentation [Ze et al, ISME J 2012]

27 Human gut microbiota Factors that impact on bacterial persistence in the colon Host factors -Microbiota acquired at birth -Host immune response Gut environmental factors nutrient availability/diet macronutrient and micronutrient availability ph bile gut transit (wash out) anaerobiosis

28 ph influences the composition of the colonic microbiota Diet (fibre/starch) Gut transit Microbial community structure Microbial activity Colonic ph Proximal colon lumen ph Main site of carbohydrate fermentation Transverse + distal colon lumen ph Available carbohydrates slowly fermentable GUT HEALTH/ GUT DISORDER Large intestinal ph

29 ph responses of two Bacteroides spp. and two butyrateproducing Firmicutes in pure culture (Duncan SH et al Environ Microbiol 2009) Bacteroides thetaiotaomicron Bacteroides vulgatus Eubacterium rectale Roseburia inulinivorans

30 Impact of ph on SCFA formation by the human colonic microbiota Short chain fatty acids - SCFA Conc. (mm) ph ph 6.5 acetate butyrate total [substrate dietary polysaccharides (mainly starch)] propionate continuous flow fermentor (Walker AW et al AEM 2005 Duncan SH et al Env Micro 2009) Time (hours) Community profile (16S rrna) E. rectale/roseburia Bacteroides

31 Short chain fatty acid metabolism - functional groups [Flint HJ et al Nat Rev GH, 2012] 1. Carbohydrate fermentors widespread - Bacteroidetes, Firmicutes, Actinobacteria Carbohydrates (hexoses) 3. Acetogens Blautia hydrogenotrophica Marvinbryantia formatixigenes (Firmicutes) 4. Methanogens Methanobrevibacter smithii (Archaea) 1 pyruvate acetyl CoA 5. Sulfate reducers Desulfovibrio piger (Proteobacteria) 2. Propionate producers Bacteroidetes Veillonellaceae CH 4 4 succinate lactate formate/ H 2 + CO 2 3 ACETATE 2 butyryl CoA 6 5 PROPIONATE SO 4 H 2 S BUTYRATE 6. Butyrate producers Faecalibacterium prausnitzii Eubacterium rectale, Roseburia spp. Eubacterium hallii, Anaerostipes spp. (Firmicutes)

32 Modelling the impact of diet on microbial fermentation in the colon* formate H 2 + CO 2 B10 CH 4 NSP oligo- starch protein saccharides, sugars B3, B6 B2, B4, B5 B1 B1-B9 PEP B3, B1 pyruvate B1-B9 acetyl CoA B9 acetate B5, B6 B8 succinate B1 lactate B7 Bacterial functional groups :- B1 = Bacteroidetes B2 = Firmicutes (eg. R. bromii) B3 = Firmicutes (eg. E. eligens) B4 = Actinobacteria B5 = Roseburia group B6 = F. prausnitzii B7 = Negativicutes B8 = E. hallii, Anaerostipes B9 = acetogens B10 = methanogens butyrate propionate (*Kettle H et al, 2014 Modelling the emergent dynamics of communities of human colonic microbiota: response to ph and peptide -

33 Modelling gut metabolism (Kettle H et al 2014) SCFA - predicted observed (Walker et al 2005) Bacterial populations - Agreement butyrate, acetate excellent propionate good succinate, formate - overestimated Dominance of B1 (Bacteroides) at ph 6.5 well reproduced based on experimentally determined ph profiles.

34 Conclusions Adequate dietary fibre may be critical in delivering long term health benefits Both the amount and type of carbohydrate (Inulin, NSP, RS) in the diet can impact on the composition of the gut microbiota and microbially-produced metabolites Carbohydrate enriched diets increased faecal concentrations of butyrate and ferulic acid derivatives compared to low carbohydrate diets Studies on isolated bacteria and consortia can provide the key information needed to model the behaviour of the system Future challenges! Understand how diet determines microbial competition and metabolic outputs from the gut microbial community using theoretical modelling Understand the causes and consequences of inter-individual variation in gut microbiota composition

35 Microbial Ecology Harry Flint (Head of group) Faith Chung Xiaolei Ze Alvaro Belenguer Shui Ping Wang Jenny Laverde Alan Walker Molecular Nutrition Wendy Russell BioSS Helen Kettle Grietje Holtrop OMH, HNU Gerald Lobley Alex Johnstone Petra Louis Freda Farquharson Karen Scott Jenny Martin Acknowledgements Collaborators: Julian Parkhill, Trevor Lawley (Sanger Institute, Cambridge, UK) Hermie Harmsen, Gjalt Welling (U. Groningen, The Netherlands) Mireia Lopez-Siles and Jesus Garcia Gill (U. Girona, Spain) Jerry Wells and Oriana Rossi (U. Wageningen, The Netherlands) George Macfarlane (U. Dundee, UK) Annick Bernalier and Christophe Chassard (INRA, Clermont Ferrand, France)